Zed2i fov - exact values

Hi,

this may be the dumbest question of the week, but what are the exact fov of the Zed2i camera ?
Here I read ZED 2 - AI Stereo Camera | Stereolabs 110° (H) x 70° (V) x 120° (D) max.

But is 110° the fov of each individual captor, or the combined fov of the two eyes ? I hope this superbly-drawn schema is explicit.

Hi @Francois
this information 110° (H) x 70° (V) x 120° (D) max. regards the raw frames provided by each single CMOS sensor.
The final FoV after image rectification is a little smaller, you can retrieve this information real-time from the CameraParameters data structure of the ZED SDK: C++/Python.
Please do not consider the global FoV created by stitching the left and the right images. The camera is designed to overlap the frames the most as possible to be able to perform the stereo-matching processing on the maximum available FoV. For this reason, a stitched image would result in an image with a FoV slightly bigger.

Great, thanks for the answer

Last question.

On a video, I have a final fov of 92.8°. Instead of the initial 110°

Is the center of the final fov aligned with the fov of the left lens ?
In that case, we “lose” 8.6° in the left of the image, and 8.6° in the right of the image.
Or is it biased and can we lose more of one side of the image ?

The rectified left image is centered in the optical center of the left lens.
The rectified right image is centered in the optical center of the right lens.
So you lose ~4.3° on the left and ~4.3° on the right of each image.

What about the depthmap ?

Each pixel of the depth map corresponds to the same pixel of the left color image, so the FoV is thecsame

Thanks for the answer on a Sunday. Does that mean that the depthmap’s FOV is 110° ?

No, the FoV of the depth map matches the value of the rectified left image, hence in your case ~93°

Ok, thanks a lot for your help

My pleasure :slightly_smiling_face:

Walter
Stereolabs Support

Just to be clear, because I have trouble getting the right values.

I get the left image

left_image = sl.Mat()
zed.retrieve_image(left_image, sl.VIEW.LEFT)

I get the depthmap

depth_image = sl.Mat()
zed.retrieve_measure(depth_image, sl.MEASURE.DEPTH)

Here, the depth_image has a fov_x=92.8 and the left_image has a fov_x=110 ?

Hi @Francois

Why do you say this?
They both have the same h_fov and the value is retrievable runtime in the way that I described here:

Ok, so that means both have fov_x = 92.8 ?

Yes, that’s what I said